As I'm working on the "filesystem on a db file" task, I need some information about how darcs uses the filesystem, to make the most efficient representation. I've read some of the source code, but I have one question for the more experienced hackers: during the apply of a patch, darcs load the whole file in memory and then writes it back, or it uses smart "offset" tricks to read and write only specific parts of the file?
In the first case, a "flat" representation of data in file is more appropriate (one data record for each file in database). In the second case a "chunked" representation of data (n data record for each file, each of the same dimension) is more efficient. Thank you, Salvatore _______________________________________________ darcs-devel mailing list darcs-devel@darcs.net http://lists.osuosl.org/mailman/listinfo/darcs-devel
